The development of artificial intelligence methods has impacted therapeutics, personalized diagnostics, drug discovery, and medical imaging. Although, in many situations, AI clinical decision-support tools may seem superior to rule-based tools, their use may result in additional challenges. Examples include the paucity of large datasets and the presence of unbalanced data (i.e., due to the low occurrence of adverse outcomes), as often seen in neonatal medicine. The most recent and impactful applications of AI in neonatal medicine are discussed in this review, highlighting future research directions relating to the neonatal population. Current AI applications tested in neonatology include tools for vital signs monitoring, disease prediction (respiratory distress syndrome, bronchopulmonary dysplasia, apnea of prematurity) and risk stratification (retinopathy of prematurity, intestinal perforation, jaundice), neurological diagnostic and prognostic support (electroencephalograms, sleep stage classification, neuroimaging), and novel image recognition technologies, which are particularly useful for prompt recognition of infections. To have these kinds of tools helping neonatologists in daily clinical practice could be something extremely revolutionary in the next future. On the other hand, it is important to recognize the limitations of AI to ensure the proper use of this technology.

Continuous glucose monitoring using subcutaneous sensors is useful in the management of glucose control in neonatal intensive care. We evaluated feasibility and reliability of a continuous glucose monitoring system in a population of very low birth weight neonates needing parenteral nutrition. Moreover, we presented percentiles of glycemia of the studied population. Very low birth weight neonates were enrolled within 24 h from birth. An ENLITE sensor connected to a continuous glucose monitoring system was inserted and maintained for at least 72 h. Data obtained with the continuous glucose monitoring system and with a glucometer were compared. Calibration was performed every 12 h. Twenty-three patients (9 males) were included. Median gestational age was 28 weeks (range 23–30) and median birth weight was 860 g (range 500–1092). A total of 299 paired glucose values were obtained. Modified Clarke Error Grid criteria for clinical significance were met. 74 and 33 episodes of hypoglycemia and hyperglycemia were detected, respectively. 31,329 values of glycemia were analyzed and the percentiles calculated. This continuous glucose monitoring system is safe and accurate. It allows increasing the detection of hypo- and hyper-glycaemia episodes and it could be routinely used in the management of glucose infusion in very low birth weight neonates under total parenteral nutrition.

(1) Background: Empirical antibiotics for suspected neonatal early-onset sepsis are often prolonged administered, even in the absence of clinical signs of infection, while awaiting the blood cultures results. The C-reactive protein is widely used to guide antibiotic therapy, although its increase in the first hours of life is not always evidence of infection. The aim of this study was to evaluate the time to positivity (TTP) of blood cultures (BC) that develop pathogens in our population of neonates and determine whether TTP could safely inform the decisions on empirical antibiotic discontinuation in neonatal early-onset sepsis and reduce the use of unnecessary antibiotics. (2) Methods: We retrospectively collected data of all newborns ≥ 34 weeks admitted to the Neonatal Intermediate-Care Unit at Policlinico “A. Gemelli” University Hospital (Rome, Italy) from 2014 to 2018, with suspected early-onset sepsis (EOS). The TTP was the time in hours from the first BC inoculation to the bacterial growth. We defined as positive BC only those with a pathogenic organism. (3) Results: In total, 103 out of 20,528 infants born in the five-year study period were admitted to our Neonatal Intermediate-Care Unit because of a suspected EOS and enrolled into the study. The mean TTP of pathogenic organisms was 17.7 ± 12.5 h versus 80.5 ± 55.8 h of contaminants (p = 0.003). We found ten positive BCs. The TTP of BC was lower than 12, 36, and 48 h in 80%, 90%, and 100% of cases, respectively. CRP levels on admission were similar in infants with a positive and negative BC (p = 0.067). The discontinuation of therapy in asymptomatic infants 48 h after initiation would have resulted in a saving of 217 days of antibiotics (31.1% of total days administered). (4) Conclusion: From our data, the TTP of blood cultures that develop pathogens is less than 48 h in 100% of cases. Therefore, in late preterm and full-term infants with suspected EOS, stopping empiric antibiotics 48 h after initiation may be a safe practice to reduce unnecessary antibiotic use, when blood cultures are negative and infants asymptomatic.

Background: Diaphragm dysfunction is associated with poor outcomes in critically ill patients. Ventilator‐induced diaphragmatic dysfunction (VIDD), including diaphragm atrophy (DA), is poorly studied in newborns. We aimed to assess VIDD and its associations in newborns.Methods: Single‐center prospective study. Diaphragm thickness was measured at end‐inspiration (TDI) and end‐expiration (TDE) on the right midaxillary line. DA was defined as decrease in TDE ≥ 10%. Daily measurements were recorded in preterm newborns on invasive mechanical ventilation (IMV) for ≥2 days. Clinical characteristics of patients and extubation failure were recorded. Univariate analysis, logistic regression, and mixed models were performed to describe VIDD and associated factors.Results: We studied 17 patients (median gestational age 27^0/7 weeks) and 22 IMV cycles (median duration 9 days). Median TDE decreased from 0.118 cm (interquartile range [IQR] 0.094–0.165) on the first IMV day to 0.104 cm (IQR 0.083–0.120) on the last IMV day (p = .092). DA occurred in 11 IMV cycles (50%) from 10 infants early during IMV (median: second IMV day). Mean airway pressure (MAP) and lung ultrasound score (LUS) on the first IMV day were significantly higher in patients who developed DA. DA was more frequent in patients with extubation failure than in those with extubation success within 7 days (83.3 vs. 33.3%, p = .038).Conclusions: DA, significantly associated with extubation failure, occurred in 58.8% of the study infants on IMV. Higher MAP and LUS at IMV start were associated with DA. Our results suggest a potential role of diaphragm ultrasound to assess DA and predict extubation failure in clinical practice.

Chest and abdominal X-rays after the insertion of an epicutaneo-caval catheter in infants are the standard method of checking the tip location in many neonatal intensive care units. The role of ultrasound in the tip location of the epicutaneo-caval catheter in neonates has been the subject of many recent studies. This systematic review investigates the accuracy of epicutaneo-caval catheter tip location by comparing ultrasound and conventional radiology. We performed a systematic literature search in multiple databases. The selection of studies yielded nineteen articles. The systematic review and meta-analysis were performed according to PRISMA (Preferred Reporting Items for Systematic reviews and Meta-analysis). The analyses showed that ultrasound is a better imaging technique for epicutaneo-caval catheter tip location in the neonatal intensive care unit than conventional radiology. By improving operator training and selecting a standardized echography protocol, ultrasound could become the gold standard for visualizing the epicutaneo-caval catheter tip in the neonatal intensive care unit. This would have some important benefits: (1) increased accuracy in tip location (2); a more rapid use of the central venous access (3); and a significant reduction in radiation exposure.